[Aldaryn]

Hi!

Had some freetime, and built a simple cornell like testing environment. I wanted to localize, and understand how black dots populate the scene.

No tweaking through the UI (shutter, ISO) or PS was done.

[n1tram]

Great test Aldaryn! I tried something like this and found a strong correlationship with black dots and UV roughness + sunlight.

I then tried a interior scene with emitters behind dielectrics (all caustic layers on) and my whole render was a big white dot

Every object was outlined by white dots, and mos of my textures where not even visible.

Any thoughts?

[Thomas An.]

Very nice test !
Thank you

[aitraaz]

Good god. Magenta dots. Anarchy

[Aldaryn]

I've attached a new image. I think all the dot related problems are mainly from multiple transmitted light. The more complex a transmitted light situation is, the more dots you get.

There is no problem with direct, and mainly diffuse indirect light. But even a dielectric seen through another generates white (or with higher abbe, colored) dots.

[Aldaryn]

I've uploaded the image to imageshack, the previous one was a hungarian image hosting service, maybe a bit slow from abroad.

[Aldaryn]

Btw, has anyone notieced how sss brightens up the whole image? This needs further testing...

[Hervé]

thanks.... well it kinda gives more bightness to the sss surfaced objects, that is why I think we see it brighter...

[Becco_UK]

Aldaryn: That is an impressive amount of testing you have done. At the very least it informs users what combinations to avoid until the problem is fixed.

Have you noticed the 'pink' dots are virtually the same colour as the garish pinkish purple in the Maxwell Render spectrum gradient. Might there be a connection?

[Hervé]

the now famous garish pinkish.... hehe

[tom]

thank you for your valuable time aldaryn! nice test!

[aitraaz]

Ah Ah..."While this solution worked, it was fundamentally flawed, creating the otherwise contradictory systemic anomaly, that, if left unchecked, might threaten the system itself."...

[Aldaryn]

Becco_UK: I dont think there is a straight connection between the pink in the spectrum and the doths here, however, there might be some.
In this case, it might just be a result of the color from scattering objects.

Another indirect situation:


Mutiple specular reflected light hitting a scattering plane, and not a single black dot. Seem so, dots mostly appear in multiple transmitted light situations.

The black dots on the white plane are just noise. The preview showed no dots at level 14.